Vous êtes sur la page 1sur 5

ISSN 1648-2824 KALB STUDIJOS. 2002. Nr. 2 * STUDIES ABOUT LANGUAGES. 2002. No.

A Comparison of Reading Models, Their Application to the Classroom and Their


Impact on Comprehension
Vida kudien
Abstract. Reading, an important skill in any language, becomes more difficult in a foreign language. The
investigation of expository text comprehension between language groups conducted in 1999 at Vytautas
Magnus University and South Carolina University (kudien, 1999) showed that the students of both language
groups (Lithuanian and English) performed better in literal comprehension tasks. On these grounds, the
assumption was made that teachers emphasize a bottom-up rather than a top-down model of reading. The aim
of the research was to investigate whether a top-down or a bottom-up model of reading is emphasized during
pre-, while-, and post- reading activities at the intermediate level of the English language instruction. The
results of the research exhibited that most of the pre-reading and while-reading activities used in reading
instruction are based on top-down models while post-reading instruction is interactive with more emphasis on
bottom-up models.
Introduction

READING MODELS

A reading model is theory of what is going on in the


readers eyes and mind during reading and comprehending
(or miscomprehending) a text (Davies, 1995:59). Models
of the reading process try to explain and predict reading
behaviour. They are the bases on which reading
instructions are built. The proposed study focuses on the
analysis of two main models of reading: bottom-up and
top-down.

BOTTOM-UP

TOP-DOWN

APPLICATION
PRE-

WHILE-

POST-

Figure 1. The research design

The results of the previous research (kudien, 1999)


showed that the students performed better literal
comprehension tasks. Therefore, an assumption was made
that teachers emphasize bottom-up rather than top-down
models of reading.

Traditionally, comprehension has been viewed as a


readers capacity to replicate a text (e.g. in summaries of
facts, translations, or matching exercises). However, in
recent years considerable research effort has focused on a
conceptual model. According to Beaugrande, what readers
comprehend is not sentences but conceptual content
(Swaffar, 1991:39). A bottom-up model, which focuses on
linguistic clues, builds literal comprehension of a text and
a top-down model, which emphasizes the importance of
background knowledge, builds global comprehension
(Carrell, 1988; Swaffar, 1991).

Method
Instrumentation
In order to obtain the necessary data on reading models
applied by teachers a questionnaire was devised. The
structure of the questionnaire was formed by the Fraenkel
and Wallen (1993) study. It involved three multiple-choice
questions on pre-reading, while-reading and post-reading
(see Appendix) activities. The question on pre-reading
activities gave 5 possible answers; the question on whilereading activities gave 7 possible answers one of which
was divided into 3 possible sub-answers; the question on
post-reading activities included 7 answers, one of which
had 2 possible sub-answers. There was a supply-type item
after each question which gave a respondent an
opportunity to add other answers. The activities given in
the multiple-choice questions were adapted from the books
by Carrell, Devine, and Eskey (1988) Interactive
Approaches to Second Language Reading and by
Aebersold and Field (1997) From Reader to Reading
Teacher. Each activity emphasized either a bottom-up or a
top-down process of reading. The pre-reading question was
constructed of 2 activities based on a bottom-up approach
and 3 activities based on top-down models; the whilereading question involved 5 activities based on bottom-up
and 4 activities on top-down models; the post-reading

Top-down models of reading can hardly be used at


elementary levels of language instruction because, as
Carrell and Coady argue, knowledge of a minimum of
5000 words is essential to make top-down processing
possible (Swaffar, 1991:44). In contrast, bottom-up
models are not useful at the advanced levels because
students are able to decode graphical input automatically.
Reading instruction at the intermediate level is more
complex because both models are to be applied.
The aim of the research was to investigate whether a topdown or a bottom-up model of reading is emphasized
during pre-, while-, and post-reading activities at
intermediate level of the English language instruction
(Figure 1).

94

TA1P Inform students about the topic of the passage.


TA2P Give tasks, which invite the comparison between
features of the students culture and the target culture.
TA4P Invite students participation.

question included 4 activities based on bottom-up and 4


on top-down models. The total numbers of both possible
bottom-up and top-down answers were equal (418).
Sample
The questionnaire was administered to 38 teachers of an
intermediate level of English instruction. The participants
were from different schools and universities. None of them
had followed courses specifically focused on reading
comprehension teaching.
Procedures
The teachers were contacted in Vytautas Magnus
University during the conference in May, 2000. They were
asked to volunteer in the study by filling out a
questionnaire. The questionnaire was anonymous and there
were no time limits.

The data demonstrate that more teachers focus on topdown processes of reading (F=62) which, as it was
mentioned above, builds global comprehension of a text
rather than a bottom-up process of reading (F=39) which
develops literal comprehension of a text.
While-reading bottom-up activities
The second question of the questionnaire was about the
activities used during reading a text. BA1W, BA2W,
BA6W, BA7AW and BA7CW exercises were based on
bottom-up view to reading process. The eight teachers use
none of these activities. The total of circled bottom-up
activities was 50. Activity BA1W was chosen by 13
teachers, BA2W by 11, BA6W by 5, BA7AW by 9
and BA7CW by 12 teachers. The frequency and
percentage of total responses to while-reading bottom-up
activities are presented in Table 3.

Results
Descriptive statistics were conducted to show: frequency
and percentage of total responses to the tasks in terms of
two types (bottom-up and top-down) of reading models.
Statistical analysis was executed on an IBM mainframe
computer at Vytautas Magnus University using the SPSS
package of statistical analysis programs for descriptive
statistics.

Table 3. Frequency and percentage of total responses to whilereading bottom-up activities


Types of activities
BA1W
BA2W
BA6W
BA7AW
BA7CW
Total

Pre-reading bottom-up activities


'The possible answers to questions 3 and 5 in the multiple
choice-section on pre-reading activities were based on
bottom-up models of reading. The results showed that 10
teachers out of 38 use none of these activities. The bottomup activities were circled 39 times by the remaining 28
teachers. BA3P (bottom-up pre-reading activity) activity
was chosen by 20 teachers, BA5P by 19 teachers. The
frequency and percentage of total responses to pre-reading
bottom-up activities are presented in Table 1.

Frequency
20
19
39

Focus on students pronunciation, discrimination


between sounds, intonation.
BA2W Go through a text word-for-word concentrating on
unknown words.
BA6W Interrupt the reading process to explain grammatical
units.
BA7AW Translate unknown words.
BA7CW Ask students to look unknown words up in the
dictionary.

Percentage
51.3
48.7
100

While-reading top-down activities


TA3W, TA4W, TA5W and TA7BW exercises were based
on top-down models. They were circled 80 times. All of
the teachers applied at least one of them to their classroom.
Exercise TA3W was marked by 11 teachers, TA4W by
15, TA5W by 23 and TA7BW by 31 teacher. The
frequency and percentage of total responses to whilereading top-down activities are presented in Table 4.

BA3P Prepare students for linguistic difficulties in the text.


BA5P Give word-recognition/phrase-identification tasks.

Pre-reading top-down activities


The three possible choices of activities based on top-down
models were circled 62 times by ten teachers. TA1P
activity is used by 19 teachers, TA2P by 13 and TA4P
by 30 teachers.
The frequency and percentage of total responses to prereading top-down activities are presented in Table 2.

Table 4. Frequency and percentage of total responses to whilereading top-down activities


Types of activities
TA3W
TA4W
TA5W
TA7BW
Total

Table 2. Frequency and percentage of total responses to prereading top-down activities


Types of activities
TA1P
TA2P
TA4P
Total

Frequency
19
13
30
62

Percentage
26.0
22.0
10.0
18.0
24.0
100

BA1W

Table 1. Frequency and percentage of total responses to prereading bottom-up activities


Types of activities
BA3P
BA5P
Total

Frequency
13
11
5
9
12
50

Percentage
30.6
21.0
48.4
100

TA3W
TA4W

95

Frequency
11
15
23
31
80

Percentage
13.7
18.8
28.7
38.8
100

Require students to transfer information from a


continuous text to some kind of grid or matrix.
Interrupt the reading process and ask students to
predict the following events.

TA5W

Prepare students to skim by asking them to recognize


the key sentences of a passage.
TA7BW Ask students to predict the meaning of an unknown
word from the context.

The data show that, in contrast to the situation in pre- and


while-reading activities, bottom-up theory (F=86) rather
than top-down (F=84) theory is more widely applied after
reading a text.

The data show that a top-down model is applied more


often (F=80) than a bottom-up (F=50) model while reading
a text in English classrooms at the intermediate level of
language proficiency.

The difference is not significant in the number of both


kinds of activities applied. Thus, it may be said that
reading instruction after reading is almost interactive, i.e.
both bottom-up and top-down models are applied
alternatively.

Post-reading bottom-up activities


The third question in the questionnaire concerned the
activities used after reading a text. BA1A, BA3A, BA4A
and BA5AA exercises were based on bottom-up theory.
They were circled 86 times. All of the teachers used at
least one of these activities. Activity BA1A was used by 26
teachers, BA3A by 9, BA4A by 22 and BA5AA by
29 teachers. The frequency and percentage of total
responses to post-reading bottom-up activities are
presented in Table 5.

On the whole, the teachers questionnaire shows that more


attention in reading instruction is paid to top-down models
(F=226) which means that reading is basically taught as an
active, predictive process (Table 7).
Table 7. Frequency of total responses to bottom-up and top-down
activities
Type of
activities

Table 5. Frequency and percentage of total responses to postreading bottom-up activities


Types of activities
BA1A
BA3A
BA4A
BA5AA
Total

Frequency
26
9
22
29
86

Bottom-up
Top-down

Percentage
30.2
10.5
25.6
33.7
100

Give fill-in exercises.


Give tasks which require students to recognize verb
inflections, comparative forms, derivational and
other affixes.
BA4A
Ask students to memorize new words and
expressions.
BA5AA Give true/false or multiple-choice exercises in order
to enhance comprehension.

TA6A
TA7A

Total

176
226

The study aimed at assessing whether a top-down or a


bottom-up model of reading is emphasized during pre-,
while-, and post- reading activities at the intermediate level
of English language instruction.
Pre-reading activities
As far as the pre-reading activities are concerned, the
results showed that teachers emphasize top- down models
more often (F=62). Examples of these are: what activates
readers background knowledge on the topic (TA1P,
F=19), builds students cultural background (TA2P, F=13)
and encourages students to predict the events of a text to be
read (TA4P, F=30). Such kinds of methods motivate
students to read for purpose and provide a framework for
the kind of information that students will read in the text. It
emerged that the most popular pre-reading activity (TA4P,
F= 30) was to predict the events of a text.

Table 6. Frequency and percentage of total responses to postreading top-down activities

TA2A
TA5BA

Postreading
frequency
86
84

Discussion

Post-reading top-down activities


The use of activities based on top-down models is less
frequent. The four possible answers TA2A, TA5BA,
BA6A, and BA7A were circled 84 times. Each teacher
uses at least one of these activities. Activity TA2A was
used by 23 teachers, TA5BA by 29, TA6A by 26,
TA7A by 6 teachers. The frequency and percentage of
total responses to post-reading top-down activities are
presented in Table 6.

Frequency
23
29
26
6
84

Whilereading
frequency
50
80

The results of the questionnaire on the two kinds of


reading activities used before, during and after reading the
text, and the comparison in their usage were presented.
Specifically, the results reveal that in pre-reading and
while-reading activities top-down models of reading are
used more frequently than bottom-up. However, in postreading activities bottom-up models are used more
frequently than top-down.

BA1A
BA3A

Types of activities
TA2A
TA5BA
TA6A
TA7A
Total

Prereading
frequency
40
62

Percentage
27.4
34.5
31.0
7.1
100

While-reading activities
With respect to the while-reading activities, top-down
procedures are more frequently employed (F=80) than
bottom-up (F=50).

Involve students into role-play.


Conduct a discussion on the topic to enhance
students comprehension.
Ask students to write a summary or their own
opinion on the topic.
Teach students to employ mind-mapping techniques.

Some bottom-up exercises such as: BA1W (F=13) which


put emphasis on the letter-to-sound procession through a
text, BA2W (F=11) which focus on the smallest units in
constructing the meaning of a text and, as mentioned
above, destroys students chances to comprehend the main
idea of a text, BA7CW (F=12) which deals with the
explanation of the grammatical units encountered in the

96

text or looking up an unknown word in the dictionary thus


interrupting the normal reading process that may build
comprehension of separate pieces of a text and BA7AW
(F=9) which involves the translation of unknown words
during the process of reading and makes students rather
passive readers, inhibit the global comprehension of a text
and discourages students interpretations of a text.

inconsistency between
comprehension.

The overall picture that has emerged is that teachers in


many cases rely on a top-down procedure during the
reading and use such activities as: TA3W (F=11) which
encourage students to use the available clues in the
ongoing text and see how meaning accumulates throughout
the text; TA4W (F=15) which asks students to predict the
contents of the next part of a text and helps them to
develop strategies for contextual guessing; TA5W (F=23)
which relates to the skimming process as a quick and
superficial reading of a text in order to get the gist, and the
most popular activity TA7BW (F=31) which asks students
to predict the meanings of unknown words from the
context and helps them to become more active readers.

reading

instructions

BOTTOMUP

TOPDOWN

LITERAL

GLOBAL

and

Figure 2. Inconsistency between reading instruction and


comprehension

There are several possible explanations for this


inconsistency. The first one is that reading instructions
emphasizing top-down theory are not sufficiently welldeveloped and that teachers fail to teach students to use
higher-order knowledge in generating the meaning of a
text. The students acquired the skills of word recognition at
elementary levels, and they are not able to acquire the
skills of predicting about meaning at intermediate levels.
Thus, it is easier for them to simply recognize the words on
the page. When students are not able to engage
successfully in an appropriate degree of knowledge-based
processing, they overrely on text-based processes and try
to construct meaning from the textual input only.

Post-reading activities
According to the questionnaire results the post-reading
activities almost equally differentiate between the models.
However, a little bit more attention is paid to a bottom-up
(F=86) than a top-down (F=84) model. Among the most
frequently utilized bottom-up exercises are: BA1A (F=26)
which deals with fill-in exercises and asks for accurate
answers based on mechanical clues and is no guarantee of
comprehension or coherent self-expression (Swaffar,
1991:33); BA4A (F=22) which may have a negative
impact on comprehension too because when students know
that their vocabulary will be tested they are more likely to
adopt systematic textual decoding than to resort to
extensive guessing; BA5AA (F=29) which fails to teach
students to express their relative perceptions because it
asks about unconnected facts only. The above mentioned
practice helps to concentrate on recognition and recall the
so-called lower-order cognitive tasks and draws heavily on
retentive capability (Swaffar, 1991:71).

The second explanation is that the students suffer from


what Samuels and Kamil (1988) called a meaning is in
the text belief. It is common to classroom settings where
reading is often done for the teachers purposes, and where
reading comprehension is usually tested by question
answering for a grade. Students fail to correctly answer
questions about texts that require extra-textual knowledge.
When Spiro interrogated such students informally, they
appeared to be perfectly able to answer the same questions.
To the question why they did not utilize the same
knowledge to answer after reading, the students responded
that they thought they were not supposed to (Carrell and
Eisterhold, 1988:79).
Self-report measures show various methodological
limitations and the questionnaire employed in the present
study is not exempt from these limitations. For instance,
the subjective evaluation of how frequently the two models
of reading are applied might not correspond to the actual
use of both strategies. However, the consistency of the
rating patterns that emerged suggests that the instrument
managed to identify a common system of reading
instruction. Furthermore, no social desirability or
compliance effects seem to have occurred, because there is
no reason suspect that participants had distorted their
spontaneous responses in order to make them match the
expectations.

The top-down activities implemented in this mode are:


TA2A (F=23) which is the role-play type of training to be
used when appropriate background knowledge is to be
developed or activated; TA5BA (F=29) which functions to
teach students to interpret the text and, most important,
give students a chance to benefit from the thoughts and
knowledge of their classmates; TA6A (F=26) which
involves summary writing and allows students more
control over the language as well as build their global
comprehension of a particular text. TA7A (F=6) which
unfortunately is not frequently used, helps students to
select the key content from a passage and represent it in
some sort of visual display (Carrell, 1988:249).

Conclusions

It was assumed that teachers emphasize a bottom-up rather


than a top-down theory because students performed better
in literal comprehension tasks in the test made by kudien
(1999). However, the assumption was rejected: teachers
employ more often a top-down (F=226) rather than a
bottom-up (F=176) model of reading. Figure 2 shows

The aim of the research was to investigate whether a topdown or a bottom-up model of reading is emphasized
during pre-, while-, and post-reading activities at the
intermediate level of the English language instruction. The
composed questionnaire filled in by the 38 teachers of
different institutions served as the instrument of the
research. The research showed that most of the pre-reading

97

and while-reading activities used in reading instruction are


based on top-down models while post-reading instruction
is an interactive emphasizing of bottom-up models.
Therefore, the conclusion was reached that teachers see
reading as active, predictive process which in larger pieces
of a text, aids readers' comprehension. The results of the
research on the application of the activities were not
consistent with the view that top-down models build global
comprehension. Consequently, the conclusion was drawn
that the application of the activities was not sufficiently
well-developed or that students think detailed recall of the
information in the text, and not the ability to interpret it, is
needed for a good grade. The inconsistency proposes that
the effects of reading instruction on comprehension
demand further studies as well as the analysis of much
more data.

References

Taking into account the results of the study, three


suggestions for classroom applications can be made. First,
teachers should be more enthusiastic in involving students
participation in reading comprehension skill development.
Second, teachers should be provided with more theoretical
information on teaching reading. Third, students need to be
encouraged to be more active in the classroom. Finally,
teachers should take into account the primary importance
of a global comprehension skill at the intermediate level of
language instruction and novel competence.

1.

Aebersold, J.A. and Field, M.L. (1997) From Reader to Reading


Teacher. Issues and Strategies for Second Language Classrooms.
New York: Cambridge University Press.

2.

Carell, P., Devine, J. and Eskey, D. (1988) Interactive Approaches


To Second Language Reading. Cambridge: Cambridge University
Press.

3.

Carrell, P. (1988) Interactive Text Processing: Implications for


ESL/Second Language Reading Classrooms.
Interactive
Approaches To Second Language Reading. Rev.ed. Ed. Carrell, P.,
Devine, J. and Eskey, D. (p.p. 239-257). Cambridge: Cambridge
University Press.

4.

Carrell, P. and Eisterhold, J. (1988) Schema theory and ESL reading


pedagogy. Interactive Approaches To Second Language Reading.
Rev. ed. Ed. Carrell, P., Devine, J. and Eskey, D. (p.p. 73-89).
Cambridge: Cambridge University Press.

5.

Davies, F. (1995) Introducing Reading. London: Penguin.

6.

Fraenkel, J.R. and Wallen, N.E. (1993) How to Design and Evaluate
Research in Education. New York: McGraw-Hill Inc.

7.

Goodman, K.S. (1988) The reading process. Interactive Approaches


To Second Language Reading. Rev. ed. Ed. Carrell, P., Devine, J.
and Eskey, D. (p.p. 11-21). Cambridge: Cambridge University
Press.

8.

Samuels, S.J. and Kamil, M.L. (1988) Models of the reading


process. Interactive Approaches to second language reading. Rev.
ed. Ed. Carrell, P., Devine, J. and Eskey, D. (p.p. 22-35).
Cambridge: Cambridge University Press.

9.

Swaffar, J.K., Arans, K.M. and Byrnes, H. (1991) Reading For


Meaning. Integrated Approach to Language Learning. N Jersey:
Prentice Hall.

10. kudien, V. (1999) Ekspozicinio teksto supratimo mokymo ir


mokymosi efektyvinimo prielaidos dviej kalbini grupi analizs
pagrindu. Kaunas: VMU.

Vida kudien
Skaitymo modeli lyginimas, j taikymas klasje ir taka supratimui
Santrauka
Straipsnyje pristatoma dviej pagrindini skaitymo modeli apaia-virus ir virus-apaia taikymo angl kalbos vidutinio lygio grupse analiz.
Tyrimo instrumentu buvo pasirinkta anketa, sudaryta remiantis Franenkel ir Wallen (1993), Carrell (1988) ir Aebersold (1997) pateiktais pasilymais.
Konferencijos, kuri vyko Vytauto Didiojo universitete 2000 met gegus mnes metu buvo surinkti duomenys i 38 atsitiktinai parinkt mokytoj ir
dstytoj i vairi Lietuvos mokykl ir universitet.
Kaip parod tyrimo rezultatai, virus-apaia skaitymo modelis yra taikomas daniau negu apaia-virus kas leido teigti, kad mokymo procese yra
labiau akcentuojamas visuminio teksto supratimo gebjimo lavinimas. Prielaida, kad skaitymo mokymo procese yra labiau akcentuojamas apaiavirus modelis, kuris lavina paodin teksto supratim, nepasitvirtino. Remiantis tyrimo duomenimis buvo padarytos ivados: 1) visuminis skaitymo
supratimo gebjimas nra pilnai ir isamiai lavinamas, 2) studentai nesupranta visuminio teksto supratimo reikms, 3) mokytojai ir dstytojai neturi
pakankamai informacijos apie visuminio skaitymo supratimo gebjimo lavinimo svarb. Siekdami padti besimokantiesiems gyti taip reikaling
visuminio skaitymo supratimo ini, mokytojai ir dstytojai turt lavinti auktesniojo skaitymo supratimo lygmens gebjimus, mokyti auktesniojo
skaitymo lygmens strategij, pvz., kaip reikia atskirti svarbius informacinius vienetus nuo papildomos informacijos vienet, taip pat mokyti teksto
apibendrinimo strategij.
Straipsnis teiktas 2001 11
Parengtas publikuoti 2002 05
The author
Vida kudien, Assoc. Prof. Dr., International School of Management, Lithuania.
Academic interests: linguistics, skill development.
Address: International School of Management, E. Oekiens 18-101, LT-3000 Kaunas, Lithuania.
E-mail: : vidsku@ism.lt

98